This invention provides a system and method of Advanced Process Control for optimal operation of multi-unit plants in large scale processing and power generation industries. The invention framework includes the following components: continuous real time dynamic process simulation, automatic coefficient adjustment of dynamic and static process models, automatic construction of transfer functions, determination of globally optimal operating point specific to current conditions, provision of additional optimal operating scenarios through a variety of unit combinations, and calculation of operational forecasts in accordance with planned production.
Legal claims defining the scope of protection, as filed with the USPTO.
1. An advanced process control system for controlling a continuous processes comprising: a. a continuous real time dynamic process simulator running in parallel to a real continuous process; b. a Coefficient Adjustment Module that performs automatic coefficient adjustment of dynamic and static process models based on changes in the real process using a sequential Monte Carlo particle filter method; c. a multivariable predictive controller operating on a computing system; d. a real-time optimization module for receiving input from the multivariable predictive controller and real time data from a distributed control system; and e. a scheduling module on the computing system to create a schedule based upon the inputs and outputs of elements a.-e.
2. An advanced process control system of claim 1 , wherein the process is a multi-unit plant operation in large scale processing and power generation industries, wherein a process is considered a large scale when it has thousands of control variable and control actions with tens of thousands of independent variables.
3. An advanced process control system of claim 1 , wherein the continuous real time dynamic process simulator uses dynamic process models consisting of a set of differential equations.
4. An advanced process control system of claim 1 , wherein the Coefficient Adjustment Module automatically adjusts both static and dynamic process models.
5. An advanced process control system of claim 1 , wherein the Coefficient Adjustment Module uses ordinary least squares, partial least squares, decision trees, and artificial neural networks to build process models.
6. An advanced process control system of claim 1 , wherein the Coefficient Adjustment Module automatically adjusts coefficients of both static and dynamic process models based on process changes using particle filters, also known as Sequential Monte Carlo (SMC) methods.
7. An advanced process control system of claim 6 , wherein the process changes are detected by a time series and six sigma based shift detection algorithm.
8. An advanced process control system of claim 1 , wherein the multivariable predictive controller provides real time content to the Operator via an integrated Visualization Module, which includes state-of-the-art data visualization components; this provides visual representation of static and dynamic models and visual representation of optimal operating point relative to modeled operating envelope.
9. An advanced process control system of claim 1 , wherein the multivariable predictive controller predicts future changes in controlled variables and determines past changes in manipulated variables and disturbance variables.
10. An advanced process control system of claim 1 , wherein the multivariable predictive controller calculates new changes in manipulated variables in order to ensure that control variable set points are reached and account for Operator chosen optimization criteria.
11. An advanced process control system of claim 1 , wherein the real-time optimization module automatically constructs transfer functions through simulation.
12. An advanced process control system of claim 11 , wherein the transfer functions are generated by simulated open-loop step performed on the current dynamic process model.
13. An advanced process control system of claim 11 , wherein the process simulation is analytically achieved using a set of differential equations that comprise the dynamic process model.
14. An advanced process control system of claim 1 , wherein the real-time optimization module simulates a disturbance variable and manipulated variable step change test such that disturbance variables and manipulated variables are changed separately to observe the controlled variable (CV) response.
15. An advanced process control system of claim 1 , wherein the real-time optimization module performs several open-loop step change tests for each variable to obtain the transfer function within acceptable accuracy thresholds that are set by process operators or the system.
16. An advanced process control system of claim 11 , wherein transfer functions can be generated for parallel, in-series, combination, and ambient disturbance variable input/output process architectures, in a setting of a large scale multi-unit process.
17. An advanced process control system of claim 11 , wherein transfer functions are described by the first order plus time delay form, in a setting of a large scale multi-unit process.
18. An advanced process control system of claim 1 , wherein the real-time optimization module automatically determines optimal operating mode specific to current conditions.
19. An advanced process control system of claim 1 , wherein the real-time optimization module utilizes a variety of optimization techniques including integer programming, linear programming, mixed integer programming, mixed integer non-linear programming, quasi-Newton method, Nelder-Mead Simplex Method, and Lagrange multipliers.
20. An advanced process control system of claim 1 , wherein the real-time optimization module automatically provides additional optimal operating scenarios through a variety of unit combinations.
21. An advanced process control system of claim 20 , wherein the available optimal scenarios (based on a range of expected future conditions) are relayed to the Operator along with economic assessments through the Visualization Module, which includes state-of-the-art data visualization components.
22. An advanced process control system of claim 1 , wherein the scheduling module calculates operational forecasts in accordance with planned production.
23. An advanced process control system of claim 22 , wherein the planned production is provided by the Operator and can be set for any time period.
24. An advanced process control system of claim 1 , wherein the scheduling module employs genetic algorithms to find optimal solutions to efficient operating mode problems, in particular, for finding an optimal combination of unit shut-downs and start-ups in a large scale process.
25. An advanced process control system of claim 1 , wherein the scheduling module employs genetic algorithms to find optimal solutions to forecast parameter search problems.
26. An advanced process control system for controlling continuous processes comprising: a. a Coefficient Adjustment Module that performs a continuous real time dynamic process simulator running in parallel to a real process; b. a module performing an automatic coefficient adjustment of dynamic and static process models; c. a multivariable predictive controller operating on a computing system; d. a real-time optimization module for receiving input from the multivariable predictive controller and real time data from a distributed control system; e. a scheduling module on the computing system to create a schedule based upon the inputs and outputs of elements a.-e; f. wherein the Coefficient Adjustment Module automatically adjusts coefficients of both static and dynamic process models based on process changes using particle filters, also known as Sequential Monte Carlo (SMC) methods; and g. wherein the process changes are detected by a time series and six sigma based shift detection algorithm.
27. An advanced process control system for controlling continuous processes comprising: a. a continuous real time dynamic process simulator running in parallel to a real process; b. a module performing an automatic coefficient adjustment of dynamic and static process models; c. a multivariable predictive controller operating on a computing system; d. a real-time optimization module for receiving input from the multivariable predictive controller and real time data from a distributed control system; e. a scheduling module on the computing system to create a schedule based upon the inputs and outputs of elements a.-e; and f. wherein the real-time optimization module performs several open-loop step change tests for each variable to obtain the most accurate transfer function.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
May 25, 2012
April 14, 2015
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.